CAMUS_public-ImageMask-Dataset

Dataset Overview

Dataset Name

CAMUS_public-ImageMask-Dataset

Dataset Description

CAMUS_public (Cardiac Acquisitions for Multi‑structure Ultrasound Segmentation) is an image‑segmentation dataset comprising clinical examinations from 500 patients collected at the University Hospital of Saint‑Étienne, France, fully anonymized according to the hospital ethics committee. The dataset targets left‑ventricular ejection fraction measurement without any pre‑selection, preserving clinical diversity.

Dataset Composition

• Training set: 450 patients with expert‑annotated references.
• Test set: 50 new patients.
• File format: Raw input images are supplied as .raw/.mhd files.

Dataset Structure

• Raw dataset layout:

  ./CAMUS_public ├─database_nifti │ ├─patient0001 │ ├─patient0002 │ … │ └─patient0500 └─jupyter

  Each patient folder contains two file types: *.nii.gz (image) and _gt.nii.gz (mask).

• Processed dataset layout:

  ../CAMUS_public-ImageMask-Dataset ├─test │ ├─images │ └─masks ├─train │ ├─images │ └─masks └─valid ├─images └─masks

Dataset Download

• Original dataset: CAMUS_public
• Processed version: CAMUS_public-ImageMask-Dataset.zip

Citation

• Paper: S. Leclerc, E. Smistad, J. Pedrosa, A. Ostvik, et al. "Deep Learning for Segmentation using an Open Large‑Scale Dataset in 2D Echocardiography" in IEEE Transactions on Medical Imaging, vol. 38, no. 9, pp. 2198‑2210, Sept. 2019.
• DOI: 10.1109/TMI.2019.2900516